yba@mit-athena.ARPA (Mark H Levine) (06/25/84)
My understanding of AI is severely limited by lack of experience; I have just finished one introductory course, so please take these as the questions of a beginner: Why does the question of a soul arise at all? It seems completely irrelavant. As far as I know, no one has ever demonstrated by scientific or other empirical means that humans have souls--while it may be commonly accepted by some or even most people, it is certainly not a test-able fact that humans have souls. At least, no one has written out the means by which I can see for myself that I have a soul. I would, of course, be very interested in such a proof either way. Why then bother about computer (programs, systems, AI systems, &c.) souls? Murder is taking a man's life. Most people who believe in the soul seem to feel that the soul is untouched (in the victim) by the act of murder. I don't want to dwell on things beside the point. The relavance of anything to AI, I would expect, would be judged by what its presence or lack contributes to the performance or functionality of AI systems. Since we do not understand the mechanism of souls and intuition in men, it seems pointless to speculate on whether an AI system will have same. It is more to the point to try and build a mechanism called "intuition" or "soul" and see if it can successfully reproduce human thought processes or behavior. If a computer using a random number generator to produce a "canvas" on the crt produces an image which humans agree has strong emotional content (is "Art"), have we built an artistic soul? An answer to that question may reveal a great deal about how people make value judgements; it will not say anything about a soul. In my personal experience, I have found that I can make intuitive leaps; I can suspend disbelief long enough to entertain a hypothesis which flys in the face of reason; I can take something on faith (like a soul) if I "want to". All the programs I have built can not do any of these things. I like to try and let understanding cover what can be understood, and then stop. It seems to me that certain questions are unanswerable by their own definition (like do I have a soul? We define it as something intangible and then have no place to view it from--any answer must be an act of faith); these things are so much six thousand year old baggage that I do not think should be allowed to slow us down any longer. Yet the basic lack of security in all of us owing to "imponderables" seems always to present itself. Why can't we just say "We do not know and cannot find out" and then get on with it? If we can find out at some future point because the rules change (God could really exist, show up on a Tuesday, and give a proof) we can deal with it then! A long while back a fellow explained that all human suffering is caused by attachment. Anybody got a better handle on truth? I want to see a system that can teach a beginner how to keep his beginner's mind; now how do I do it? -- yba%mit-heracles@mit-mc.ARPA UUCP: decvax!mit-athena!yba